userfriendlyscience NEWS
------------------------
0.7.2 2018-09-11
* Fixed bug in plots for genlog and piecewiseRegr that caused odd Y axis labeling
* Added functionality to makeScales to play nice with Ruben Arslan's `codebook` package
* Started migrating functions to `behaviorchange` and `ufs` packages
*
0.7.1 2018-05-01
* Added dataset with aggregate scales originally published with therapy monitor
* Added genlog, for generalized logistic analysis, plus helper functions
* Added separate functions for getting starting values and plotting starting values for genlog
* Added 'lettering' of post hoc differences in oneway based on https://stackoverflow.com/questions/48280985
* Fixed compatibility issue of manual with new version of `car` package
0.7-0 2017-11-17
* Added facComAnalysis function, to approximate SPSS' FACTOR function
* Added 'reliability' function, to approximate SPSS' RELIABILITY function
* Added validMeans and validSums, to only compute aggregates when there are enough valid values
* Improved ggPie by correcting label positioning and adding argument to set colors
* Added multiResponse, to generate a table for multiple response questions
* Added ggEasyRidge, to easily generate ridgeline plots for multiple variables
* Added ggEasyBar, to easily visualise multiple categorical variables with the same categories
* Added knitFig, to generate separate chunks (with separate sizes) for figures in e.g. R Markdown
* Added multiVarFreq, the tabular equivalent of ggEasyBar
* Added logRegr for logistic analysis
* Added fanova as flexible anova wrapper for simple anova's
* Added collinearity diagnostic to regr and logRegr
* Added userfriendly plotting to regr, fanova and logRegr
* Added sharedSubString to find the longest shared substring is a character vector
* Added findShortestInterval to find the shortest interval in a numeric vector
* Added checkDataIntegrity for easy checking of data integrity
* Improved nnc to compute confidence intervals using formula from
* Added confIntProp for easily computing confidence intervals for proportions
* Added ggProportionPlot to generate a plot for proportions acknowledging inaccuracy of sample estimates
* At the request of Dorota Reis (through Twitter, @dororeis), made it easy to change the colors in dlvPlot()
* In CIBER, it's now possible to number the subquestions
0.6-1 2017-05-09
* Added support for Cohen's d diamond plots
* Added function for confidence curves
* Fixed measurement level related bug in associationMatrix
* Added processLimeSurveyDropouts
* Minor improvements in CIBER
* Added possibility to easily reset counter for caption
0.6-0 2017-04-01
* Added CIBER for Confidence Interval-Based Estimation of Relevance
* Added a number of diamond plot functions (biAxisDiamondPlot, associationsDiamondPlot) and supporting functions (varsToDiamondPlotDf, associationsToDiamondPlotDf, rawDataDiamondLayer)
* Added pwr.randomizationSuccess
* Small improvements to cohensdCI and pwr.cohensdCI
* Added confIntR and pwr.confIntR based on the Moinester & Gottfried (2014) and
the Bonett (2011) papers (with permission)
* Added determinant structure specification functions (determinantStructure, determinantVar, subdeterminants, subdeterminantProducts)
* Added determinant structure processing functions (detStructAddVarNames, detStructAddVarLabels, detStructComputeProducts, detStructComputeScales, detStructCIBER)
* Added function to easily process LimeSurvey variable lables (processLSvarLabels)
* Added curfnfinder (from Nick Sabbe, with permission)
* Added escapeRegex (from Hmisc, to prevent importing package)
* Fixed warning-generating 'bla' in pwr.cohensdCI
0.5-2 2016-12-15
* Fixed some errors in the list of exported functions
* descr function now uses the quantile function to compute the interquartile range
* Minor fix in dlvPlot (removed unknown aesthetic)
* Fix in invertItems to no longer simplify a dataframe to a vector if only one
column is selected
* Added Coefficient H to scaleStructure
* Minor improvement in getData to enable suppressing messages
* Fix in meanDiff to better interpret measurement levels of input variables
* Added pwr.cohensdCI for 'power analyses' for Cohen's d confidence intervals
* Fix in 'modus' function (for finding the mode of a vector)
0.5-1 2016-11-09
* Minor tweaks to work with newest version of ggplot2
* No longer exporting all functions using regex but explicitly naming them now
0.5-0 2016-11-07
* Power function for Omega Squared was added
* diamondPlot functions were added
* Pander methods for some functions were added
* Some functions to make R more accessible for SPSS users were added (e.g. regr, oneway, examine)
* Confidence intervals implemented in Tukey and Games-Howell posthoc tests\
* Simple interfaces for generating ggplot scatterplots, Q-Q plots, boxplots, barcharts and pie charts were added
* showPearsonPower, a function to show the power of a Pearson correlation was added
* Cohen's d and Omega Squared distributions were added by converting to and from t and F
* A function to detect rare words was added
* Functions to automate figure and table captions in R Markdown/knitr were added
* therapyMonitor was added, a function to aid n-of-1 analyses for therapists
* prevalencePower was added, a function for power analyses for prevalence studies
0.4-0 2015-
* processOpenSesameIAT was added, to process OpenSesame IAT datafiles
0.3-0 2015-04-16
* Fixed powerHist to finally show frequencies with scaled normal curve
and density curve
* Fixed a bug in 'oneway'; post-hoc tests now working better
* The 'regr' function now also computes the confidence interval for R^2
* Renamed scaleReliability to scaleStructure and improved it to also provide
estimates assuming an ordinal level of measurement as well as hierarchical
omega in addition to total omega
* Improved the way p values are reported in rMatrix
* Improved dlvPlot to show both confidence intervals and standard errors
0.2-2 2015-04-07
* Error in getData, that was introduced when it was changed to work better with getDat, corrected
* Way eta^2 is computed in oneway corrected
* A plot can now automatically be ordered when using meanDiff
0.2-1 2015-03-23
* getDat added to automatically create a dataframe
* Minor additions to the conversion functions
* Added and implemented formatPvalue for cleaner p value formatting
* Exported the case insensitive %IN%
0.2-0 2015-02-27
* New function associationMatrix() added, which takes a list of variables
and generates a table with confidence intervals of their effect sizes
* New functions freq(), oneway(), and regr() added, which behave similar
to SPSS functions to generate frequency tables, oneway anovas, and
do simple regression analyses
* New function samplingDistribution() added, which allows quick generation
of sampling distributions based on frequency information
* Added didacticPlot(), which builds plots that are useful for watching
what happens when doing null hypothesis significance tests
* Added dataShape() so that e1071 is no longer needed for Skewness and Kurtosis
in normalityAssessment()
* Added associationMatrix, which enables the creation of matrices with
effect sizes and their confidence intervals.
0.1-2 2014-04-21
* testRetestReliability() added to conveniently call both testRetestAlpha()
and testRetestCES().
* testRetestAlpha() and testRetestCES() added to compute the test-retest
coefficient Alpha and the Coefficient of Equivalence and Stability (CES).
* filterBy() and useAll() added for users coming from SPSS
* convertToNumeric and massConvertToNumeric added to convert vectors to
numeric vectors.
* meanDiff.multi improved to generate a forest plot by default, with line at
zero to show 'significant' deviations more clearly
* dlvPlot improved to show better confidence intervals and take confidence level
into account
* getData() can now also read from .csv, .tsv, .xls, and .xlsx. If
ROpenOffice from http://omegahat.org is installed, .ods can be read,
too.
* Added is.odd() and is.even() to check whether numbers (or numbers in a vector) are odd or even.
0.1.1 2014-04-17
* Initial Release